Clinical Care/Education/Nutrition/Psychosocial Research ORIGINAL ARTICLE

Sequential Intensification of Metformin Treatment in Type 2 Diabetes With Liraglutide Followed by Randomized Addition of Basal Insulin Prompted by A1C Targets

1 6 J. HANS DEVRIES, MD, PHD ANNE B. THOMSEN, MD, PHD etformin is generally considered to 2 6 STEPHEN C. BAIN, FRCP MARCIN ZYCHMA, MD, PHD fi 3 7 be the most appropriate rst-line HELENA W. RODBARD, MD JULIO ROSENSTOCK, MD M 4 pharmacotherapy for treating type JOCHEN SEUFERT, MD, PHD ON BEHALF OF THE LIRAGLUTIDE-DETEMIR 5 2 diabetes (1), but there is no general agree- DAVID D’ALESSIO, MD STUDY GROUP* ment on how to advance treatment when metformin becomes insufficient. With many d drug classes available, translational studies are OBJECTIVE We evaluated the addition of liraglutide to metformin in type 2 diabetes fol- needed to identify the most effective, safe, and lowed by intensification with basal insulin (detemir) if glycated hemoglobin (A1C) $7%. simplest antidiabetes treatment sequences. RESEARCH DESIGN AND METHODSdIn 988 participants from North America and Practical treatment strategies that achieve Europe uncontrolled on metformin 6 sulfonylurea, sulfonylurea was discontinued and and maintain glycated hemoglobin (A1C) lev- liraglutide 1.8 mg/day added for 12 weeks (run-in). Subsequently, those with A1C $7% were els at ,7% while minimizing hypoglycemia randomized 1:1 to 26 weeks’ open-label addition of insulin detemir to metformin + liraglutide and weight gain are especially desirable (2). n n , ( = 162) or continuation without insulin detemir ( = 161). Patients achieving A1C 7% Glucagon-like peptide 1 receptor ago- continued unchanged treatment (observational arm). The primary end point was A1C change nists (GLP-1RAs) reduce A1C by 0.8–1.5% between randomized groups. and weight by 2–3 kg on average in combi- RESULTSdOf 821 participants completing the run-in, 61% (n = 498) achieved A1C ,7% nation with metformin (depending on study (mean change 21.3% from 7.7% at start), whereas 39% (n = 323) did not (20.6% from 8.3% populations and other background therapy), at start). During run-in, 167 of 988 (17%) withdrew; 46% of these due to gastrointestinal and are associated with a low risk of hypogly- adverse events. At week 26, A1C decreased further, by 0.5% (from 7.6% at randomization) cemia (3–6). Although the durability of effec- with insulin detemir (n = 162) versus 0.02% increase without insulin detemir (n = 157) to 7.1 tiveness of GLP-1RAs has not yet been 2 2 2 P , and 7.5%, respectively (estimated treatment difference 0.52 [95% CI 0.68 to 0.36]; established, most other type 2 diabetes drugs 0.0001). Forty-three percent of participants with insulin detemir versus 17% without reached , fail after several years, such that many patients A1C 7%. Mean weight decreased by 3.5 kg during run-in, then by 0.16 kg with insulin eventually require insulin treatment to attain detemir or 0.95 kg without insulin detemir. In the randomized phase, no major hypoglycemia occurred and minor hypoglycemia rates were 0.286 and 0.029 events per participant-year and sustain glycemic control. This is typically with and without insulin detemir (9.2 vs. 1.3%). achieved by adding a basal insulin to previous medications because such insulins are associ- CONCLUSIONSdSupplementation of metformin with liraglutide and then insulin detemir ated with modest hypoglycemia risk and was well tolerated in the majority of patients, with good glycemic control, sustained weight loss, weight gain compared with premixed and and very low hypoglycemia rates. prandial insulins (7,8). The side effects of in- – sulin might be mitigated if used with a GLP- Diabetes Care 35:1446 1454, 2012 1RA; however, to date, there have been only trials adding GLP-1RAs to insulin and no well-controlled trials examining basal insulin added to existing GLP-1RA therapy. In this ccccccccccccccccccccccccccccccccccccccccccccccccc trial, for the first time, we evaluated a novel From the 1Department of Internal Medicine, Academic Medical Centre, Amsterdam, the Netherlands; the fi 2 treatment intensi cation sequence: adding a Institute of Life Sciences, Swansea University and Abertawe Bro Morgannwg University NHS Trust, GLP-1RA (liraglutide) to metformin followed Swansea, U.K.; 3Endocrine and Metabolic Associates, Rockville, Maryland; the 4Division of Endocrinology and Diabetology, University Hospital of Freiburg, Freiburg, Germany; the 5Division of Endocrinology, by a randomized, open-label investigation of Diabetes, and Metabolism, University of Cincinnati, Cincinnati VA Medical Center, Cincinnati, Ohio; further intensification with systematically ti- 6Medical & Science, GLP-1 & Obesity, Novo Nordisk A/S, Soeborg, Denmark; and the 7Dallas Diabetes and trated basal insulin (insulin detemir) in par- Endocrine Center at Medical City, Dallas, Texas. ticipants with $7% A1C. Corresponding author: Julio Rosenstock, [email protected]. Received 4 October 2011 and accepted 28 February 2012. DOI: 10.2337/dc11-1928. Clinical trial reg. no. NCT00856986, clinicaltrials.gov. RESEARCH DESIGN AND This article contains Supplementary Data online at http://care.diabetesjournals.org/lookup/suppl/doi:10 METHODS .2337/dc11-1928/-/DC1. *A complete list of the members of the Liraglutide-Detemir Study Group can be found in Supplementary Table 13. © 2012 by the American Diabetes Association. Readers may use this article as long as the work is properly Participants cited, the use is educational and not for profit, and the work is not altered. See http://creativecommons.org/ Eligible participants were insulin-naïve licenses/by-nc-nd/3.0/ for details. adults (18–80 years) with type 2 diabetes

1446 DIABETES CARE, VOLUME 35, JULY 2012 care.diabetesjournals.org DeVries and Associates treated for $3monthswith$1,500 Assessments and end points per group were needed to detect a 0.5% mg/day metformin and A1C values of The primary end point was the change in A1C between-group difference with 90% 7.0–10.0% or with metformin and sul- A1C (%) from randomization (week 0) to power. fonylurea (less than or equal to half of week 26 to determine whether adding Unless stated otherwise, efficacy end the maximum approved dose) and A1C insulin detemir to metformin + liraglutide points were analyzed statistically using values of 7.0–8.5% (Supplementary was superior to continued metformin + the full analysis set (all randomized par- Data online). liraglutide. Additional efficacy end points ticipants with greater than or equal to included the following: percentage of one efficacy value) with missing values Trial design and interventions participants reaching A1C ,7and imputed by carrying forward the last The trial was conducted in 202 office- or #6.5%, FPG, postprandial plasma glu- observation. An ANCOVA model with hospital-based sites in Belgium, Canada, cose from self-measured seven-point glu- treatment, country, and previous oral France, Germany, Italy, the Netherlands, cose profiles, weight, blood pressure, lipids, antidiabetic therapy as fixed effects and Spain, the U.K., and the U.S. between 3 and percentage of participants reaching the randomization value as covariate was March 2009 and 19 April 2010. Protocol composite end point of A1C ,7% with no used to analyze change from randomiza- amendments occurring after the study weight gain or hypoglycemia (during the tion to week 26. Percentages of partic- start are summarized in the Supplemen- 26-week period). Safety assessments in- ipants reaching ,7and#6.5% A1C were tary Data online. Protocol, amendments, cluded adverse events (AEs) and hypogly- analyzed using logistic regression with and informed consent documents were cemic episodes. Minor hypoglycemic the same fixed effects and covariate. A approved by independent local ethics episodes (plasma glucose ,3.1 mmol/L) similar logistic regression model was committees and implemented according were self-treated. Major episodes required used for the composite end point (pro- to good clinical practice (9) and the third-party assistance, irrespective of portion reaching ,7% A1C with no Declaration of Helsinki (10). plasma glucose levels. weight increase and no hypoglycemia This 38-week, open-label trial com- during the 26-week period). The safety prised a 12-week run-in followed by a Statistical analyses analysis set comprised all participants ex- 26-week, randomized, two-armed, parallel- For the primary end point, superiority of posed to at least one dose of trial drug. group period for participants not achieving the insulin detemir over the control group Hypoglycemic episodes were analyzed ,7% A1C. At run-in start, sulfonylurea was was concluded if the 95% CI upper limit using a generalized linear model. Descrip- discontinued (in approximately one-third of for the treatment difference was ,0. As- tive statistics were calculated for the obser- participants), and liraglutide was initiated in suming 20% run-in withdrawals and 40% vational group. Run-in results were 0.6-mg/day weekly increments to a final randomization eligibility, 150 participants summarized by the randomized period 1.8-mg/day dose (Fig. 1A). The metformin dose remained unchanged. Participants with A1C $7% at the end of run-in were randomized (1:1) to 26 weeks’ insulin detemir added to met- formin + liraglutide 1.8 mg (randomized detemir group) or continued metformin + liraglutide 1.8 mg only (randomized con- trol group) (see Supplementary Data online for randomization details). A double- blind, placebo-controlled approach was not used because titration with detemir placebo was not feasible. To determine the effects of continued treatment on ini- tial responders, participants with A1C ,7% after run-in were followed for 26 weeks as a prespecified observational group. Insulin detemir (100 units/mL) and liraglutide (6.0 mg/mL) (Novo Nordisk A/S, Bagsvaerd, Denmark) were injected sub- cutaneously once daily with pen devices. Insulin detemir was administered with evening meals or at bedtime; liraglutide was administered at any (consistent) time of day. Insulin detemir treatment d fl was started at 10 units and titrated by Figure 1 Trial design (A) and trial ow diagram (B). *Two participants who had been ran- domized to the metformin and liraglutide control group received the wrong trial treatment. One investigators on a weekly basis using a fi – participant was supplied with insulin detemir but withdrew before administering the treatment. speci c algorithm targeting 4.1 6.0 The other participant should not have been randomized as her A1C level at week 0 was ,7%. For mmol/L self-measured fasting plasma the full analysis, these participants appear in the randomized control group, and for the safety glucose (FPG) values (Supplementary analysis, they appear in their respective “treatment” groups (i.e., insulin detemir and observa- Data online). tional groups, respectively). care.diabetesjournals.org DIABETES CARE, VOLUME 35, JULY 2012 1447 Sequential intensification of metformin

with metformin only before enrollment (Table 1). At run-in completion, A1C was reducedby1.3%inthisobservationalgroup and by 0.6% in the randomized groups from A1C values at run-in start of 7.7 and 8.3%, respectively (Fig. 2A and Supplementary Table 1); weight decreased by 3.5–4.4 kg (Fig. 2B)andFPGby1.0–2.0 mmol/L (Fig. 2C). One major hypoglycemic event occurred (blood glucose level, 5.2 mmol/L) and minor hypoglycemia occurred at 0.000–0.372 events per participant-year (Supplementary Table 2). Nausea was the most frequently reported run-in AE (Table 2) but incidences fell to ,7% in all groups after 3 weeks (Supplementary Table 3 and Supplemen- tary Fig. 1). One case of acute pancreatitis was reported. In another participant, with elevated calcitonin level (23.5 ng/L) before liraglutide administration, a subsequent thyroidectomy revealed an incidental thyroid neoplasm (1-mm papillary micro- carcinoma) (Supplementary Table 4).

Adding insulin detemir to metformin and liraglutide in participants requiring additional glycemic control: the 26-week, randomized period Participants not reaching glycemic tar- get had a 7.6% mean A1C at completion of the run-in period. They were then ran- domized to add insulin detemir to metfor- min + liraglutide (n =162)ortocontinue unchanged metformin + liraglutide (ran- domized control subjects, n = 161). The groups’ characteristics were similar at run- in (Table 1). Addition of insulin detemir further reduced A1C compared with continued metformin + liraglutide (20.51% [n =162] vs. +0.02% [n = 157], respectively; esti- mated treatment difference [ETD] 20.52 [95% CI 20.68 to 20.36]; P , 0.0001), resulting in A1C values of 7.1 and 7.5%, respectively, at week 26 (Fig. 2A and Sup- plementary Tables 5 and 6), and meeting the trial’s primary end point. Accordingly, Figure 1dContinued mean FPG decreased more in the detemir (22.1 mmol/L) than control group (20.4 mmol/L; ETD 21.7 [22.2 to 21.3]; P , treatment group (i.e., observational, ran- (46%) of these withdrew due to gastroin- 0.0001) (Fig. 2C). After a mean 3.5-kg domized detemir, or randomized control), testinal AEs (7.7% of enrolled partici- weight loss during run-in, both groups ex- although all participants received the same pants). Of 821 participants completing perienced further modest weight reduction run-in treatment. the run-in, 498 (61%) reached ,7% over the next 26 weeks: 20.16 kg with A1C; 323 (39%) did not and were eligible detemir and 20.95 kg without (ETD 0.79 RESULTS for randomization. [0.08–1.49]; P = 0.03) (Fig. 2B). Compared with participants that did Compared with the control group, Adding liraglutide to metformin: not reach the target during run-in, those more than twice as many in the insulin 12-week run-in reaching the target had a shorter type 2 detemir group achieved ,7% A1C (17 vs. During the run-in, 167 of 988 (17%) diabetes duration and lower A1C and 43%, respectively; P , 0.0001), and three participants withdrew (Fig. 1B); 76 of 167 FPG values and more had been treated times as many achieved #6.5% (6 vs. 18%,

1448 DIABETES CARE, VOLUME 35, JULY 2012 care.diabetesjournals.org DeVries and Associates

Table 1dDemographics and disease characteristics at run-in start (week 212) and randomization (week 0)

Randomized treatment group Randomized control group Observational group metformin + liraglutide 1.8 mg + metformin + liraglutide metformin + liraglutide insulin detemir (n = 162) 1.8 mg (n = 161) 1.8 mg (n = 498) At run-in (week 212) Age (years) 56.8 (9.4) 57.3 (9.8) 56.5 (9.7) Duration of diabetes (years) 8.6 (5.8) 8.5 (6.0) 6.6 (5.7) Male:female (%) 54.3:45.7 55.3:44.7 56.6:43.4 Previous oral antidiabetic drug metformin:metformin + sulfonylurea (%) 50.0:50.0 50.3:49.7 74.5:25.5 BMI (kg/m2) 34.9 (6.3) 33.9 (6.0) 34.4 (6.7) Weight (kg) 99.5 (21.2) 98.6 (21.3) 99.0 (20.8) A1C (%) 8.2 (0.7) 8.3 (0.8) 7.7 (0.7) HOMA-B 59.0 (50.8) 51.2 (34.9) 63.7 (46.1) Fasting plasma glucose (mmol/L) 10.2 (2.4) 10.3 (2.5) 9.2 (1.8) Systolic blood pressure (mmHg) 134.0 (16.9) 135.7 (16.8) 134.4 (15.3) Diastolic blood pressure (mmHg) 80.1 (9.7) 80.8 (9.8) 81.5 (9.2) At randomization (week 0) Weight (kg) 96.0 (20.9) 95.3 (21.1) 94.7 (20.5) A1C (%) 7.6 (0.6) 7.6 (0.7) 6.4 (0.4) Fasting plasma glucose (mmol/L) 9.2 (1.9) 8.8 (2.1) 7.2 (1.3) Systolic blood pressure (mmHg) 132.2 (16.3) 131.7 (14.9) 128.9 (15.2) Diastolic blood pressure (mmHg) 79.6 (9.8) 80.9 (9.4) 79.4 (9.5) Data are means (SD) unless otherwise noted. Data exclude participants that withdrew during the run-in period prior to randomization. HOMA-B, homeostasis model assessment of b-cell function. respectively; P = 0.0016; logistic regression There were no significant differences end of run-in were followed for another estimates) (Supplementary Fig. 2). Simi- between groups in changes from random- 26 weeks. This group experienced a mean larly, the proportion reaching the compos- ization for serum lipids, except for free 1.3% A1C reduction by the end of run-in, ite end point (,7% A1C with no weight fatty acids (insulin detemir group, 20.11 and 1.1% overall from start of run-in to gain and no hypoglycemia) after 26 weeks mmol/L; control group, 20.03 mmol/L; week 26 (from 7.7% at run-in start to was significantly greater in the insulin ETD 20.08 [95% CI 20.13 to 20.03]; 6.6% at study end). The group also expe- detemir (21%) than the control group (9%; P = 0.002) (for more information regard- rienced overall reductions from start of P = 0.0016; logistic regression estimates) ing changes in lipid levels, see Supple- run-in to week 26 in FPG of 2.1 mmol/L (Supplementary Fig. 2). Excluding one mentary Table 7). (from 9.2 mmol/L at run-in start to 7.2 outlier in the randomized control group A total of 67% (109 of 163) and 59% mmol/L) and weight of 4.8 kg (from with 25 minor hypoglycemic episodes, mi- (93 of 159) of participants had one or more 99.0 kg at run-in start to 94.6 kg) (Fig. nor hypoglycemia rates were 0.286 and AEs, and 5.5% (9 participants; 13 events) 2), and improved seven-point blood glu- 0.029 events per participant-year for insu- and 3.8% (6 participants; 8 events) had cose profiles (completers) (Fig. 2D). lin detemir and control groups, respectively serious AEs (SAEs), in the insulin detemir AEs for the 38-week period are sum- (P = 0.004) (Supplementary Table 2). No and control groups, respectively (Table 2). marized in Supplementary Table 10. In major hypoglycemic events occurred dur- No pattern or clustering of SAEs was ob- the observational group, 81% (402 of ing the randomized period. served, with most being considered unlikely 499) of participants had AEs and 7.8% Self-measured plasma glucose levels to be related to treatment (Supplementary (39 of 499) experienced 49 SAEs, with decreasedinbothgroups,withsignificantly Table 8). Although more AEs of increased 45 considered unlikely to be related to greater reductions in postprandial values lipase were reported with insulin detemir, a study drug and without obvious pattern in the insulin detemir versus control group minor increase in the median serum lipase (Supplementary Table 8). No major hy- (Fig. 2D). ETDs ranged from 20.60 mmol/L levels (below the upper limit of normal) was poglycemic episodes occurred, whereas (95% CI 21.12 to 20.08; P =0.02)to21.12 observed across all groups, without appar- 9.0% (45 of 499) of participants experi- mmol/L (95% CI 21.72 to 20.51; P = ent treatment differences (Supplementary enced minor hypoglycemia (0.211 events 0.0003). Mean prescribed insulin detemir Table 9). One case of chronic pancreatitis per participant-year). See Supplementary doses increased from 10 units/day to 39.5 occurred (control group) (Supplementary Table 9 for additional safety assessments. units/day (0.41 units/kg), whereas mean Table 4). self-measured FPG decreased from 7.9 to Systolic blood pressure and heart 6.2 mmol/L (Fig. 2E). Prescribed insulin Efficacy and safety over 38 weeks: rate: run-in to 26 weeks detemir doses were consistent with algo- the observational group At 26 weeks, systolic blood pressure was rithm recommendations, indicating good The 498 participants achieving the ,7% reduced from run-in start by 3.13 mmHg overall adherence to the titration algorithm. target with metformin + liraglutide at the in the control group, 1.65 mmHg in the care.diabetesjournals.org DIABETES CARE, VOLUME 35, JULY 2012 1449 Sequential intensification of metformin

Figure 2dGlycemic efficacy, changes in body weight, and insulin detemir doses. Results from run-in to randomization (vertical dotted line) and from randomization to the end of the trial for change in A1C (A), change in body weight (B), and mean FPG (C). Data are means 6 2SEfromthefull analysis set with no imputation. RT, randomized treatment group; RC, randomized control group; O, observational group. D: Self-monitored plasma glucose profiles before and after breakfast, lunch, and dinner and at bedtime for the treatment and control groups at weeks 0 and 26.-, randomized treatment group; ▫, randomized control group; 4, observational group. Dotted lines, 0 weeks; solid lines, 26 weeks. Vertical bars indicate 6 2SEM. P values refer to differences between groups in the change from randomization (week 0) to week 26. E: Insulin detemir dose (prescribed dose, ○, left ordinate) and self-measured FPG (◆, right ordinate) during the 26-week randomized period for the insulin detemir treatment group. Dotted lines indicate 25th–75th percentiles.

1450 DIABETES CARE, VOLUME 35, JULY 2012 care.diabetesjournals.org DeVries and Associates

Figure 2dContinued insulin detemir group, and 3.33 mmHg in treatment, 61% of run-in completers the hypoglycemia risk was very low, and the observational group (Supplementary reached an A1C ,7%. For the 39% of insulin detemir recipients maintained Table 11). Heart rate was increased from completers not initially achieving target, the significant weight reductions achieved run-in start to 26 weeks in all treatment adding insulin detemir provided clini- during liraglutide run-in rather than groups: by 3.62 beats per minute (bpm) cally relevant further reductions of mean gaining weight as typically occurs with in the control group, 3.87 bpm in the in- A1C from 7.6 to 7.1%. A total of 43% of insulin. sulin detemir group, and 3.64 bpm in insulin detemir recipients then achieved Traditional clinical practice initiates the observational group (Supplementary the ,7% A1C target, compared with 17% insulin when multiple oral therapies are Table 12). continuing treatment with metformin + no longer effective. However, after met- liraglutide. In the observational group, formin, commonly used oral agents such CONCLUSIONSdThis large, pro- all of whom achieved A1C ,7% after as sulfonylureas and thiazolidinediones spective clinical study provides support the 12-week run-in, 74% remained at tar- carry side effects, particularly hypoglyce- for a new sequential treatment paradigm in get after a further 26 weeks’ treatment. mia and/or weight gain; the exceptions type 2 diabetes, that is, adding a GLP-1RA Translated into clinical practice, and extrap- are dipeptidyl peptidase-4 inhibitors and (liraglutide) to metformin followed (if nec- olating from all patients who completed a-glucosidase inhibitors, which are essary) by a basal insulin (insulin detemir) run-in, sequential intensification enabled weight-neutral but appear to show less to achieve and maintain glycemic targets. approximately three-quarters of patients to glucose-lowering activity than other With 12 weeks of liraglutide + metformin achieve an A1C ,7%. Importantly, overall, agents (1,11). Because GLP-1RAs provide care.diabetesjournals.org DIABETES CARE, VOLUME 35, JULY 2012 1451 Sequential intensification of metformin

Table 2dAEs during run-in and randomized period with an incidence ‡5% by system organ class and preferred term and summary of SAEs

Randomized treatment Randomized Observational Early group (metformin + control group group withdrawals liraglutide 1.8 mg + (metformin + (metformin + (metformin + liraglutide insulin detemir) liraglutide 1.8 mg) liraglutide 1.8 mg) 1.8 mg) n % En% En% En% E Safety analysis set (N) 163 159 499 166 Run-in period (week 212 to 0) Total AEs 84 51.5 218 90 56.5 236 320 64.1 975 122 73.5 383 SAEs* 1 0.6 1 3 1.9 3 13 2.6 14 5 3.0 5 Nausea 23 14.1 30 31 19.5 38 120 24.0 172 66 39.8 72 Vomiting 9 5.5 11 11 6.9 11 40 8.0 61 33 19.9 42 Diarrhea 11 6.7 11 14 8.8 15 49 9.8 69 21 12.7 25 Dyspepsia 5 3.1 6 6 3.8 6 31 6.2 35 11 6.6 14 Upper abdominal pain 1 0.6 1 2 1.3 2 9 1.8 12 9 5.4 14 Headache 12 7.4 17 12 7.5 17 43 8.6 62 13 7.8 22 Lipase increased 6 3.7 6 8 5.0 8 18 3.6 18 6 3.6 7 Asthenia 3 1.8 3 d 5 1.0 5 11 6.6 13 Decreased appetite 11 6.7 11 7 4.4 7 46 9.2 47 17 10.2 17 Randomized period (week 0 to 26) Total AEs 109 66.9 413 93 58.5 323 295 59.1 790 N/A SAEs* 9 5.5 13 6 3.8 8 27 5.4 35 N/A Nasopharyngitis 23 14.1 25 30 18.9 39 45 9.0 54 N/A Diarrhea 19 11.7 26 11 6.9 12 19 3.8 19 N/A Nausea 6 3.7 6 9 5.7 10 14 2.8 16 N/A Lipase increased 18 11.0 18 6 3.8 7 17 3.4 17 N/A Headache 10 6.1 15 13 8.2 20 27 5.4 43 N/A n, number of participants with AE; %, proportion of participants in analysis set having AEs; E, number of AEs; N/A, not applicable. *A listing of all SAEs is provided in Supplementary Table 11. greater glycemic efficacy and weight loss despite an increased insulin requirement) is difficult in the absence of an active com- (6), they may be more beneficial than (15), initiating liraglutide before detemir parator or a placebo group (i.e., a group these oral agents after metformin, but may further reduce the hypoglycemia not receiving liraglutide). Elevated lipase less acceptable in some patients because risk, perhaps by modulating endogenous (or amylase) levels by themselves are in- of gastrointestinal side effects and the in- glucose-dependent insulin secretion and sufficient to make a diagnosis of pancrea- jection barrier. Insulin initiation also usu- glucagon secretion as well as body weight, titis without the presence of abdominal ally causes weight gain and an increased thereby lowering exogenous insulin re- pain and ideally confirmation with imag- hypoglycemia risk (1,12). In this study, quirements. ing studies. Most commonly, no gastroin- however, the insulin detemir group main- It remains unknown whether an as- testinal symptoms were associated with tained the initial weight loss achieved by sociation exists between incretin-based fluctuations in lipase observed during adding liraglutide to metformin. Subse- therapies and pancreatitis (16), but acute this trial. quent intensification with insulin detemir pancreatitis is known to be about three We found no other published, pro- allowed 43% of the group to reach the times more common in patients with spective, controlled trials studying insulin glycemic target. It is conceivable that type 2 diabetes than in the general popu- added to a GLP-1RA in type 2 diabetes. A more aggressive insulin titration, giving lation (17). In the present trial, two cases few previous studies have investigated the higher insulin doses, would allow even of pancreatitis were reported (one acute alternative intensification sequence, more patients to reach A1C target, but and one chronic) and neither case of pan- adding a GLP-1RA (specifically, twice- perhaps at the expense of more hypogly- creatitis was preceded by elevated lipase daily exenatide) to insulin (19,20–22). In cemia. The incidence of confirmed minor levels. Reports of pancreatitis were simi- the only randomized, controlled trial to hypoglycemia with insulin detemir in our larly uncommon in other liraglutide stud- date, patients with more advanced disease study, 0.286 events per participant-year, ies, although imbalances in pancreatitis at baseline (compared with the current was considerably lower than the 1.3–3.67 are noted for liraglutide and other incretin- study) already receiving 0.5 units/kg of events per participant-year in previous based therapies (16,18). An initial insulin glargine were randomized to ei- trials with insulin detemir added to oral small increase in median serum lipase ther add-on exenatide or continuation of agents (13,14). Although withdrawing below the upper limit of normal was ob- basal insulin only (19). The insulin doses sulfonylureas (as in the run-in) is associ- served across groups in the current study, were optimized in both groups by further ated with decreased hypoglycemia risks and substantial fluctuations occurred rigorous titration postrandomization. In (3.2 vs. 1.6 events per participant-year, over time. Interpreting these changes this setting, add-on exenatide with insulin

1452 DIABETES CARE, VOLUME 35, JULY 2012 care.diabetesjournals.org DeVries and Associates optimization was associated with a 1.7% It is possible that some participants were and has received grant support from Amylin; reduction in A1C levels (from a baseline disappointed with their treatment alloca- is a consultant for Ethicon, Merck Sharp & of 8.3–6.7% vs. 1.0% A1C reduction with tion and withdrew, an effect that might Dohme, Novo Nordisk, and Takeda; and has received grant support from Johnson & Johnson, basal insulin optimization only), modest not have occurred with an active compar- fi weight loss, and no significant increase in ator. It would be informative, however, if MannKind Corporation, and sano -aventis. A.B.T. and M.Z. are full-time employees of risk of minor hypoglycemia. In other non- future trials adhered to the comparative Novo Nordisk A/S. J.R. is an advisory board randomized studies, A1C reductions effectiveness concept (25). Future studies member and consulant for and has received – ranged from 0.0 to 0.9% (20 22). In the might also be of a longer duration and grant support from Amylin, Boehringer Ingelheim, current study, however, adding a GLP-1RA consider the overall costs of diabetes care Bristol-Myers Squibb, Daiichi Sankyo, Merck before basal insulin was also effective, of- (particularly the balance between the Sharp & Dohme, Eli Lilly & Co., GlaxoSmithKline, fering significantly improved glycemic con- higher prices of new medications versus Johnson & Johnson, Novartis, Novo Nordisk, trol, substantial weight loss, and a very low any potential for reduced costs associated Pfizer, sanofi-aventis, Roche, Intarcia, and Takeda hypoglycemia rate. This intensification se- with diabetes complications and less self- and has received grant support from AstraZeneca. fl quence may therefore be preferable pro- monitoring of blood glucose). A 26-week No other potential con icts of interest relevant vided that potential gastrointestinal AEs extension to the current study has been to this article were reported. The funder participated in trial design and can be accepted initially. It may also be undertaken, primarily to assess longer-term collection, review, and analysis of data. Liraglutide preferable to initiate with a longer-acting treatment safety. and insulin detemir are proprietary compounds once-daily GLP-1RA, such as liraglutide, In conclusion, in type 2 diabetes with manufactured and marketed by Novo Nordisk rather than a twice-daily shorter-acting inadequateglycemiccontrolon metformin 6 A/S. GLP-1RA to facilitate better compliance. sulfonylurea, the addition of liraglutide J.H.D., S.C.B., H.W.R., J.S., D.D., and J.R. We acknowledge that our trial has was effective and safe, enabling the majority planned the trial, determined the trial design, some limitations. More participants may of participants to reach A1C levels ,7%, and analyzed and interpreted data. A.B.T. have reached glycemic targets with a with sustained weight loss and low hypogly- implemented the study, oversaw the conduct lower FPG target for insulin titration cemia risk. For those not reaching glycemic of the trial, and cooperated with the medical (23). As some participants randomized to targets, intensification with insulin detemir writers and the trial statistician regarding sta- tistical analysis, data management, and writing continued liraglutide + metformin subse- provided clinically relevant additional the study report. M.Z. planned the trial and quently reached target without dose ad- glycemic control, sustaining previous weight determined the trial design, implemented the justments, it is also conceivable that fewer loss and the very low hypoglycemia risk. study, and analyzed and interpreted data. All patients would have needed insulin with a authors participated in drafting the manu- longer run-in period. Two factors may script and all have approved the final version. have influenced the outcome of the run- AcknowledgmentsdThis study was funded J.R. is the guarantor of this work and, as such, in period. First, the study used the higher by Novo Nordisk. J.H.D. is an advisory board had full access to all the data in the study and (1.8 mg) of the two liraglutide doses. The member for Eli Lilly & Co.; has received grant takes responsibility for the integrity of the data higher dose may have favored efficacy support from and has attended speakers’ and the accuracy of the data analysis. outcomes and weight loss but adversely bureau for Medtronic; is an advisory board Parts of this study were presented at the 71st member and has attended speakers’ bureau for Scientific Sessions of the American Diabetes affected tolerability (11,24). Potentially, Association, San Diego, California, 24–28 June fewer patients may have withdrawn early Novo Nordisk; has received grant support from Novartis; has attended speakers’ bureau for 2011, and in poster form at the Diabetes UK An- due to gastrointestinal AEs if they had Roche; and is an advisory board member, has nual Professional Conference, London, U.K., 30 been permitted to return to the 1.2-mg attended speakers’ bureau, and has received March–1 April 2011. dose or if upward titration was conducted grantsupportfromsanofi-aventis. S.C.B. has The authors gratefully acknowledge the more slowly. Second, one-third of partic- received lecture fees and research, educa- contribution of the investigators and their staff ipants were taking a sulfonylurea before tional, and clinical grants from Boehringer and of the participants of this trial. The authors the study; this was discontinued at the Ingelheim, Eli Lilly & Co., GlaxoSmithKline, thank Dr. Tulay T. Cushman, Dr. Jen Faleska, start of run-in and may have negatively Merck Sharp & Dohme, Novartis, Novo Dr.JohnSmith,andDr.HelleFrobøse fi fi influenced improvements in glycemic Nordisk, P zer, sano -aventis, Schering-Plough, Sørensen for medical writing support; Rie Elvang Servier, and Takeda. H.W.R. is a consultant Søndergaard for statistical support; and Mia control but favored weight loss. However, Skettrup Fitsios for statistical programming neither factor should have affected the 26- for Abbott Laboratories, GlaxoSmithKline, MannKind Corporation, and Takeda Phar- support in the preparation of this manuscript week randomized-group comparisons. maceuticals America; has attended speakers’ (all from Novo Nordisk). The authors acknowl- Moreover, discontinuation of sulfonylureas bureau for Amylin Pharmaceuticals, Bristol- edge Watermeadow Medical (Witney, U.K.), increases the trial’s translational validity Myers Squibb, and Eli Lilly & Co.; is a con- supported by Novo Nordisk, for providing because many physicians add an injectable sultant for and has attended speakers’ bureau writing and editing support. therapy only after failure of more than one for AstraZeneca and Merck Sharp & Dohme; oral agent, and then tend to withdraw is a consultant for and has received grant sup- sulfonylureas. We recognize that the ran- port from Biodel, Merck, and TolerX; has re- domized control group did not have a ceived grant support from MacroGenics; and is a References consultant for, has received grant support from, 1. Nathan DM, Buse JB, Davidson MB, et al.; masked placebo or active comparator; ’ this was mainly due to logistic limitations and has attended speakers bureau for Novo American Diabetes Association; European Nordisk and sanofi-aventis. J.S. has attended Association for the Study of Diabetes. with insulin comparators. However, we speakers’ bureau for AstraZeneca, Bayer, Berlin Medical management of hyperglycaemia in also chose not to include an active com- Chemie, Bristol-Myers Squibb, Eli Lilly & Co., type 2 diabetes mellitus: a consensus algo- parator to retain the focus on the efficacy GlaxoSmithKline, LifeScan, Merck Sharp & rithm for the initiation and adjustment of and safety of adding a basal insulin to a Dohme, Novartis, Novo Nordisk, Pfizer, sanofi- therapy: a consensus statement from the GLP-1RA in combination with metformin. aventis, and Takeda. D.D. is a consultant for American Diabetes Association and the care.diabetesjournals.org DIABETES CARE, VOLUME 35, JULY 2012 1453 Sequential intensification of metformin

European Association for the Study of Di- Efficacy/E6_R1/Step4/E6_R1__Guideline. pancreatitis and biliary disease observed abetes. Diabetologia 2009;52:17–30 pdf. Accessed 9 August 2011 in patients with type 2 diabetes: a retro- 2. American Diabetes Association. Standards 10. World Medical Association. Declaration spective cohort study. Diabetes Care of medical care in diabetesd2010. Di- of Helsinki. Ethical Principles for Medical 2009;32:834–838 abetes Care 2010;33(Suppl. 1):S11–S61 Research Involving Human Subjects [In- 18. Parks M, Rosebraugh C. Weighing risks 3. Buse JB, Rosenstock J, Sesti G, et al.; ternet], 2008. Available from http://www. and benefits of liraglutidedthe FDA’sre- LEAD-6 Study Group. Liraglutide once a wma.net/en/30publications/10policies/ view of a new antidiabetic therapy. N Engl day versus exenatide twice a day for type 2 b3/index.html. Accessed 9 August 2011 J Med 2010;362:774–777 diabetes: a 26-week randomised, parallel- 11. Pratley RE, Nauck M, Bailey T, et al.; 19. Buse JB, Bergenstal RM, Glass LC, et al. group, multinational, open-label trial 1860-LIRA-DPP-4 Study Group. Liraglutide Use of twice-daily exenatide in basal insulin- (LEAD-6). Lancet 2009;374:39–47 versus sitagliptin for patients with type treated patients with type 2 diabetes: a ran- 4. Nauck MA, Frid A, Hermansen K, et al.; 2 diabetes who did not have adequate domized, controlled trial. Ann Intern Med LEAD-2 Study Group. Efficacy and safety glycaemic control with metformin: a 26- 2011;154:103–112 comparison of liraglutide, glimepiride, week, randomised, parallel-group, open- 20. Samarasinghe YP, Shlomowitz A, Munro and placebo, all in combination with label trial. Lancet 2010;375:1447–1456 N, Feher MD. The short term effects of metformin, in type 2 diabetes: the LEAD 12. Holman RR, Thorne KI, Farmer AJ, et al.; exenatide therapy in insulin treated pa- (liraglutide effect and action in diabetes)-2 4-T Study Group. Addition of biphasic, tients with type 2 diabetes mellitus–anew study. Diabetes Care 2009;32:84–90 prandial, or basal insulin to oral therapy in option to achieve weight loss (Abstract). 5. Russell-Jones D, Vaag A, Schmitz O, type 2 diabetes. N Engl J Med 2007;357: Diabetologia 2008;51(Suppl. 1):S353 et al.; Liraglutide Effect and Action in 1716–1730 21. Sheffield CA, Kane MP, Busch RS, Bakst Diabetes 5 (LEAD-5) met+SU Study Group. 13. Hermansen K, Davies M, Derezinski T, G, Abelseth JM, Hamilton RA. Safety and Liraglutide vs insulin glargine and placebo Martinez Ravn G, Clauson P, Home P. A efficacy of exenatide in combination in combination with metformin and sulfo- 26-week, randomized, parallel, treat-to-target with insulin in patients with type 2 di- nylurea therapy in type 2 diabetes mellitus trial comparing insulin detemir with NPH abetes mellitus. Endocr Pract 2008;14: (LEAD-5 met+SU): a randomized con- insulin as add-on therapy to oral glucose- 285–292 trolled trial. Diabetologia 2009;52:2046– lowering drugs in insulin-naive people 22. Viswanathan P, Chaudhuri A, Bhatia R, 2055 with type 2 diabetes. Diabetes Care 2006; Al-Atrash F, Mohanty P, Dandona P. Ex- 6. DeFronzo RA, Ratner RE, Han J, Kim DD, 29:1269–1274 enatide therapy in obese patients with Fineman MS, Baron AD. Effects of ex- 14. Philis-Tsimikas A, Charpentier G, Clauson type 2 diabetes mellitus treated with in- enatide (exendin-4) on glycemic control P, Ravn GM, Roberts VL, Thorsteinsson B. sulin. Endocr Pract 2007;13:444–450 and weight over 30 weeks in metformin- Comparison of once-daily insulin detemir 23. Blonde L, Merilainen M, Karwe V, Raskin P; treated patients with type 2 diabetes. Di- with NPH insulin added to a regimen of TITRATE Study Group. Patient-directed abetes Care 2005;28:1092–1100 oral antidiabetic drugs in poorly controlled titration for achieving glycaemic goals 7. Holman RR, Farmer AJ, Davies MJ, et al.; type 2 diabetes. Clin Ther 2006;28:1569– using a once-daily basal insulin analogue: 4-T Study Group. Three-year efficacy of 1581 an assessment of two different fasting complex insulin regimens in type 2 di- 15. Swinnen SG, Dain MP, Mauricio D, plasma glucose targetsdthe TITRATE abetes. N Engl J Med 2009;361:1736–1747 DeVries JH, Hoekstra JB, Holleman F. study. Diabetes Obes Metab 2009;11:623– 8. Bretzel RG, Nuber U, Landgraf W, Owens Continuation versus discontinuation of 631 DR, Bradley C, Linn T. Once-daily basal insulin secretagogues when initiating in- 24. Zinman B, Gerich J, Buse JB, et al.; LEAD-4 insulin glargine versus thrice-daily pran- sulin in type 2 diabetes. Diabetes Obes Study Investigators. Efficacy and safety of dial insulin lispro in people with type 2 Metab 2010;12:923–925 the human glucagon-like peptide-1 analog diabetes on oral hypoglycaemic agents 16. Raschi E, Piccinni C, Poluzzi E, liraglutide in combination with metformin (APOLLO): an open randomised con- Marchesini G, De Ponti F. The association and thiazolidinedione in patients with type trolled trial. Lancet 2008;371:1073–1084 of pancreatitis with antidiabetic drug use: 2 diabetes (LEAD-4 Met+TZD). Diabetes 9. International Conference on Harmonisation. gaining insight through the FDA phar- Care 2009;32:1224–1230 ICH Harmonised Tripartite Guideline. Good macovigilance database. Acta Diabetol. 19 25. Nathan DM. Time for clinically relevant Clinical Practice [Internet], 1996. Avail- October 2011 [Epub ahead of print] comparative effectiveness studies in type 2 able from http://www.ich.org/fileadmin/ 17. Noel RA, Braun DK, Patterson RE, diabetes. Ann Intern Med 2011;154:131– Public_Web_Site/ICH_Products/Guidelines/ Bloomgren GL. Increased risk of acute 132

1454 DIABETES CARE, VOLUME 35, JULY 2012 care.diabetesjournals.org